Clarify links to code and checkpoints

README.md

@@ -7,25 +7,23 @@ task_categories:
 
 # Generative Embedding Benchmark
 
-<div align="center">
-  <a href="https://huggingface.co/datasets/LimitedMouse/Generative-Embedding-Benchmark"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Dataset-blue" alt="Hugging Face Dataset"></a>
-  <a href="https://github.com/LimitedMouse/Generative-Embedding-Benchmark"><img src="https://img.shields.io/badge/GitHub-Code-lightgrey" alt="GitHub Code"></a>
-</div>
-
 This repository publishes the fixed membership and provenance metadata for the
 Generative Embedding Benchmark (GEB). GEB contains 1,800 development questions
-and 900 held-out test questions, balanced across natural images, scene text, and
-visual documents.
+and 900 held-out test questions spanning natural images, scene text, and visual
+documents.
 
 GEB evaluates how much answer-relevant visual information a dense embedding
-makes accessible to a generative decoder. See the
-[code repository](https://github.com/LimitedMouse/Generative-Embedding-Benchmark)
-for installation instructions and the evaluation pipeline.
-
-The dataset does not redistribute source images, questions, or answers. The GEB
-task definitions load the eight original public datasets through `lmms-eval` and
-filter them with `dev.json` or `test.json`. Original datasets remain subject to
-their own licenses and access requirements.
+makes accessible to a generative decoder. The
+[GitHub repository](https://github.com/LimitedMouse/Generative-Embedding-Benchmark)
+is the main entry point for installation, training, and evaluation. Official
+Qwen3-VL-Embedding-2B/8B decoder weights are provided in the
+[checkpoint repository](https://huggingface.co/LimitedMouse/Generative-Embedding-Benchmark-Checkpoints).
+
+This dataset repository does not redistribute source images, questions,
+answers, or decoder training data. GEB task definitions load the eight original
+public evaluation datasets through `lmms-eval` and filter them with `dev.json`
+or `test.json`. The source datasets remain subject to their own licenses and
+access requirements.
 
 ## Files